ColabFold
ColabFold predicts protein folding structures and complexes using MMseqs2, which gives faster performance than AlphaFold.
Availability
Version |
Delta |
|---|---|
1.5.5 |
X |
The ColabFold databases are avaiable on Delta at:
/work/hdd/data/colabfold
The colabfold module may be loaded to set the environement variable COLABFOLD_PATH to the above path.
module load colabfold/1.5.5
MMseqs2
MMseqs2 is available on Delta as a module:
$ module avail mmseqs2
------------------------------ /sw/user/modules ------------------------------
mmseqs2/cpu.2025.09.29.postRel18 mmseqs2/gpu.2025.10.09.postRel18 (D)
D: Default Module
$ module load mmseqs2/<your_preferred_version>
There’s a version available for running on the CPU nodes that supports MPI, and there’s a version that was compiled for use with the A40 and A100 GPU nodes, which does not support MPI. The module names specify the type of node supported, the date on which the code was downloaded from the repository, and the fact that the code was downloaded after the stated release version (for example release version 18).
References
ColabFold code on GitHub: https://github.com/sokrypton/ColabFold
MMseqs2 code on GitHub: https://github.com/soedinglab/MMseqs2
MMseqs2 user guide: https://github.com/soedinglab/MMseqs2/wiki
ColabFold article: https://www.nature.com/articles/s41592-022-01488-1